This study investigated the role of cognitive interventions on word problem solving (WPS) accuracy in children who are English learners (ELs) and at risk for math difficulties (MD). Grade 3 children (n = 158) within classrooms were randomly assigned to either an untreated control group or one of three treatment conditions: paraphrase question (Restate), paraphrase relevant information (Relevant), and paraphrase all propositions (Complete). Hierarchical linear modeling showed that children who are ELs and at risk for MD had lower scores on WPS accuracy than their peers. The Relevant and Complete treatment conditions improved WPS accuracy when compared with the control condition. Additionally, the findings suggested that Relevant treatment condition enhances WPS accuracy for children who are at risk for MD.